ERIC Number: ED326577
Record Type: RIE
Publication Date: 1990-Oct
Reference Count: N/A
Comparing and Contrasting Neural Net Solutions to Classical Statistical Solutions.
Van Nelson, C.; Neff, Kathryn J.
Data from two studies in which subjects were classified as successful or unsuccessful were analyzed using neural net technology after being analyzed with a linear regression function. Data were obtained from admission records of 201 students admitted to undergraduate and 285 students admitted to graduate programs. Data included grade point averages, admission test scores, grades, gender, class rank, course instructor, and course grades. The neural net models used were the Adeline model and the Layer Back-propagation model. The neural net model makes no assumptions about the underlying distribution of the observations. In general, a neural network may be defined as a non-programmed information reduction system that develops processing abilities in response to its environment. The function of the neural network is to learn from examples. The study findings indicate that the results obtained using the neural net models are comparable with, but not the same as, those of the classical statistical approach. While the neural net technology is not a replacement for the classical techniques, it may represent a viable alternative when certain assumptions of the statistical model are grossly violated. (TJH)
Descriptors: Admission Criteria, Artificial Intelligence, Class Rank, College Entrance Examinations, Comparative Analysis, Equations (Mathematics), Failure, Grade Point Average, Grades (Scholastic), Graduate Students, Higher Education, Mathematical Models, Regression (Statistics), Sex Differences, Success, Undergraduate Students
Publication Type: Reports - Research; Speeches/Meeting Papers
Education Level: N/A
Authoring Institution: N/A
Identifiers: Neural Net Models
Note: Paper presented at the Annual Meeting of the Midwestern Educational Research Association (Chicago, IL, October 19, 1990).